Theses Doctoral

Tregs that accumulate in the encephalomyocarditis virus-infected mouse brain: Origin, compartmentalization, function, and gene signature

Puhr, Sarah

It is well recognized that regulatory T cells (Tregs) are immunosuppressive, by which they prevent systemic autoimmunity throughout life. Beyond this stereotypical function, however, a growing body of evidence demonstrates that Tregs in distinct tissues, including the visceral adipose tissue, dystrophic muscle, the flu-infected lung, and wounded skin can acquire unique functions directed by their local environment. Tregs in these tissues can employ a wide variety of mechanisms to accumulate and acquire tissue-specific function, including conversion from conventional T cells, canonical T cell receptor (TCR)-dependent expansion and non-canonical, TCR-independent, cytokine-dependent expansion. Intriguingly, the niche-specific function of tissue Tregs can be independent of, and mutually exclusive of, their immunosuppressive capacity. Together, this recent literature reveals that Tregs can accumulate in discrete tissue sites through non-canonical mechanisms, and in response to niche-specific cues can acquire distinct functions, which distinguish them from their peripheral, lymphoid Treg counterparts. Other tissue Treg populations remain to be identified and characterized. Moreover, it is unknown whether other tissue Tregs rely on non-canonical mechanisms of accumulation, and exhibit functions distinct from the typical Treg immunosuppressive role.
Tregs are known to accumulate in the CNS during infection, injury and inflammation. The CNS is an organ with distinctive architecture that maintains a regulated interaction with the peripheral immune system due to its critical function and poor regenerative capacity. While it is known that Tregs broadly protect against excessive tissue pathology in the diseased CNS, the origin, localization, function, mechanism of accumulation, and gene signature of CNS-infiltrating Tregs have not been studied, likely due to the challenge of isolating these rare cells and distinguishing them from circulating cells left over after perfusion.
Here, we establish a safe model of CNS infection using encephalomyocarditis virus and employ a series of methods to locate, monitor and isolate CNS-infiltrating Tregs free from contamination from the circulation. We show that a distinct population of thymus-derived Tregs accumulates within the cerebrospinal fluid (CSF) of the EMCV-infected CNS, independently of lymph node priming. Tregs function in this unique niche to limit excessive tissue pathology. While CNS Tregs maintain expression of core Treg signature genes, including FoxP3, their global transcriptome is more similar to that of conventional T cells (Tcons) harvested from the infected CNS than to that of peripheral Tregs. Bioinformatics analysis reveals that genes shared by CNS Tcons and CNS Tregs are also shared by Tregs and Tcons from injured muscle and from the visceral adipose tissue of aged mice, indicating that tissue inflammation and injury, rather than viral infection per se, contribute to CNS Treg accumulation, function and phenotype.
Additionally, we observe that CNS Treg accumulation during infection is associated with a simultaneous increase in meningeal/choroid plexus dendritic cells (m/chDCs), which are professional antigen presenting cells that localize to the gates of
the CNS. Splenic cDC and peripheral lymphoid Treg homeostasis are linked, and both populations can be artificially increased by treatment with the DC-poietin and adjuvant, Ftlt3L. Therefore, we hypothesized that CNS Tregs and m/chDCs may also be linked and could also be manipulated by Flt3L treatment. Indeed, treatment with Flt3L in conjunction with EMCV infection results in enhanced CNS Treg and m/chDC accumulation, independent of Flt3 receptor expression on Tregs. In an effort to determine if dendritic cells mediate CNS Treg increase during infection, we turned to a DC-ablative mouse model in which all CD11c-expressing cells express the catalytic subunit of diphtheria toxin and are depleted. Surprisingly, while splenic cDCs are completely abrogated in these mice, a portion of m/chDCs persists, unaffected. Moreover, CNS Tregs accumulate normally in these mice during infection. This data suggests an unappreciated heterogeneity in m/chDCs, and indicates that those that remain unaffected in these mice may mediate CNS Treg accumulation during infection. While characterizing m/chDC heterogeneity, we found that m/chDCs comprise three distinct subsets with unknown potential. Whereas m/chDCs were previously considered to be a homogeneous, CD45hiB220-CD11c+MHCII+ population, we have found them to contain three subsets, distinguishable by IRF8 and FcR-γ expression. This finding paves the way for further study of the origin, localization, and division of labor between these three m/chDC subsets.
In summary, our studies clarify the distinct compartmentalization, lymph node-independent accumulation, and inflammation-associated gene signature of CNS Tregs. Most importantly, these findings have implications for neuro-immune cross-talk, particularly at the interface of the CSF and brain parenchyma. That is, neural progenitors extend their apical domains into the CSF of the ventricles, and therefore may be subject to regulation by CSF-borne Tregs. Further, while many studies have focused on the differences between tissue Treg subsets, we find a core set of genes expressed by CNS Tregs, injured muscle Tregs and VAT Tregs. This data suggests that common mechanisms may be used for therapeutic manipulation of these cells.


  • thumnail for Puhr_columbia_0054D_13799.pdf Puhr_columbia_0054D_13799.pdf application/pdf 7.47 MB Download File

More About This Work

Academic Units
Microbiology, Immunology, and Infection
Thesis Advisors
Liu, Kang
Ph.D., Columbia University
Published Here
February 28, 2017